the Now & the Next

A bi-weekly speculative fiction suggesting the shape of things to come.
(sourced from trustworthy trade pubs, think tanks + frontier science news)

12,808 Signals Tracked
7 Collisions Curated
14 Industries
60+ Publications
2 weeks Signal Window

Seven structural forces converged in the last two weeks — and none of them are isolated. From the Strait of Hormuz to music publishing archives to orbital power beams, the world is being reorganized around contested resources: energy, catalogs, talent, biology, supply chains. What follows is what the signals suggest.

01
The Hormuz Multiplier

When a 21-mile strait rewrites every supply chain in the world

+

The Strait of Hormuz — 21 miles wide, 20% of the world's traded oil supply — has become the chokepoint around which every corporate supply chain strategy is now being rewritten. Iran launched a massive missile and drone attack on US Navy destroyers on May 8. The UAE accused Iran of a drone strike on an ADNOC tanker. A Chinese-owned tanker was attacked on May 4. Rare crossings — a Japanese supertanker carrying 2 million barrels, an ADNOC LNG carrier — are making headlines because they are exceptions. Meanwhile, Kearney's 2026 Reshoring Index reveals the paradox: despite record tariff pressure and reshoring investment, manufactured goods imports are still rising. Supply chain disruption is proving resistant to policy alone. P&G disclosed a $150M hit. Apple announced tariff refunds will be reinvested into US manufacturing. The dual-track strategy — optimize existing global flows while building redundancy for a world where those flows can be severed in hours — is now table stakes.

⚡ The Now

Every industrial enterprise is running a contingency that became the main scenario. The companies caught flat-footed are the ones that modeled Hormuz closure as tail risk. The ones ahead planned for it as operational reality. The board-level question has shifted from 'how do we optimize cost?' to 'which nodes can we not afford to lose?' Reshoring, nearshoring, and supply network redundancy are now capital allocation priorities — not regulatory compliance exercises.

→ What's Next

The Hormuz crisis is doing something tariffs alone couldn't: forcing permanent supply chain architecture decisions at the highest levels of every major industrial enterprise. Companies that build resilient, flexible supply networks in 2026–2028 will operate from a structural advantage in the 2030s. Those that wait for the conflict to resolve will be optimizing for a world that no longer exists. The manufacturing geography of the next decade is being drafted right now — with robotics and AI-powered automation as the economics that make domestic production viable at scale.

Marine Insight
Iran's IRGC Navy conducted a large-scale missile and drone attack on US Navy destroyers near the Strait of Hormuz on May 8, 2026. The attack involved both anti-ship ballistic missiles and drone swarms.
Marine Insight
The UAE accused Iran of a drone attack on the ADNOC tanker M.V. Barakah in the Strait of Hormuz, framing it as a direct threat to global energy security.
Marine Insight
The Chinese-owned oil tanker JV Innovation was attacked near the Strait of Hormuz on May 4, 2026 — the first attack on a Chinese-flagged vessel since the US-Iran conflict began.
Material Handling & Logistics
Kearney's 2026 Reshoring Index finds that despite record investment and the highest tariff regime in a century, manufactured goods imports are still rising — reshoring is not materializing at the promised pace.
Supply Chain Dive
P&G disclosed a $150M after-tax cost impact for fiscal year 2026 driven by increased commodity and energy costs tied to the Iran war and Hormuz disruptions.
Computerworld
Apple announced that tariff refunds will be reinvested into US innovation and advanced manufacturing — new investments beyond what was already committed.
02
The Catalog Wars

Every music archive just became an AI training asset

+

The music industry realized this fortnight that it is sitting on the scarcest asset in the AI economy: licensed, high-quality, human-created intellectual property with clear legal provenance. Avex Music Group launched a $100M fund to acquire publishing catalogs. Concord acquired Mothership Music Publishing and its 5,000-song catalog. Sony Music Group entered exclusive talks for a $4B acquisition of Recognition Music Group. Warner Music Group — having already settled with AI music generator Suno and established a licensing deal — was named to TIME's 100 Most Influential Companies specifically for its approach to AI. Tencent Music took down 250,000 songs and reviewed 600,000+ high-risk copyright cases in 2025. The window for training AI on uncleared music is closing, and everyone with a catalog now understands they hold a key to an industry's future.

⚡ The Now

Music catalogs are being repriced in real time as AI training data — not just as royalty-generating assets. Every major deal in the past two weeks has an AI provenance dimension: who owns the clean, licensed, commercially rich catalog controls the terms under which AI companies can use human creative history. The labels that moved first on licensing deals — Warner, Universal — now have structural leverage that will compound for years.

→ What's Next

What is happening in music is the leading indicator for every creative industry. Film libraries, book archives, journalism archives, photography collections — all will be repriced through this lens in 2026–2027. The AI companies that secure licensed training data now will have defensible moats. The ones that don't will face litigation, takedowns, and negotiating leverage that only grows as AI-generated content crowds out unlicensed training paths. Catalog valuation models built before 2025 are obsolete.

Music Business Worldwide
Avex launched a $100M fund as the first phase of its plan to acquire premier music publishing catalogs, companies, and culturally significant copyrights — striking its first deal immediately.
Music Business Worldwide
Concord acquired Mothership Music Publishing including over 5,000 copyrights and ongoing agreements with active songwriters, deepening its catalog holdings.
Music Business Worldwide
Sony Music Group entered exclusive negotiations to acquire Recognition Music Group from Blackstone for $3.5–4B through its joint venture with Singapore's sovereign wealth fund.
Music Business Worldwide
Warner was recognized specifically for settling its Suno lawsuit and establishing a licensing agreement — framed as a template for how rights holders should engage with generative AI.
Music Business Worldwide
Tencent Music deployed AI detection tools to identify and remove over 250,000 songs and review hundreds of thousands of high-risk copyright cases tied to AI-generated content.
03
The Grid Knows Your Name

AI's energy hunger is becoming the defining infrastructure story of the decade

+

AI data centers are no longer a tech infrastructure story. They are a national energy policy story, a real estate story, a climate story, and an orbital strategy story — simultaneously. In Texas, a developer pivoted 700 acquired acres from chemical manufacturing to a behind-the-meter power strategy after grid interconnection costs hit $35M. In Brazil, Elea Data Centers is building Rio AI City — a 3.2GW renewable-powered AI campus. Meta struck agreements with space-based solar startup Overview Energy for up to 1 gigawatt of power beamed from orbit. Star Catcher is building a power-beaming satellite constellation specifically for orbital data centers. And AI data centers are now exhibiting power demand patterns so variable they are challenging fundamental grid stability assumptions — causing utilities and regulators to rethink interconnection standards in real time. xAI reactivated a dormant Mississippi gas plant to power its Grok training infrastructure. The energy constraint is structural, not temporary.

⚡ The Now

The question is no longer what AI can do — it's where you can find the gigawatt to run it. Geographic arbitrage on energy access is becoming the defining competitive variable for AI infrastructure investment. Data center developers are acquiring power assets before building anything else. The companies securing energy first — through nuclear deals, behind-the-meter builds, long-duration storage, or space-based power — are determining where the next decade of AI compute happens.

→ What's Next

'Where is your data center?' is being replaced by 'Where is your gigawatt?' The energy constraint will shape which regions lead in AI infrastructure the same way fiber access shaped which cities led in internet infrastructure in the 1990s. Orbital solar power, which sounds like science fiction, now has a $1B purchase agreement from Meta. The companies treating energy as a strategic asset in 2026 will have infrastructure advantages that are difficult to replicate in five years.

Data Center Knowledge
A developer acquired 700 acres in Texas for AI compute and pivoted to behind-the-meter power after grid interconnection costs reached $35M — pursuing micro-nuclear and off-grid solutions.
Data Center Knowledge
Meta struck agreements with Overview Energy and Noon Energy for up to 1GW each of space-based solar power and ultra-long-duration energy storage to address its data center energy gap.
Space News
Star Catcher is developing a constellation of optical power-beaming satellites designed to transmit solar energy to spacecraft in orbit, enabling data centers in space.
Data Center Knowledge
Elea Data Centers is developing Rio AI City in Brazil — a large-scale AI data center project with up to 3.2 GW of renewable energy capacity targeting the global AI infrastructure market.
Data Center Knowledge
AI data centers exhibit highly variable, unpredictable power demands that are challenging traditional grid stability assumptions, forcing regulators and utilities to rethink interconnection standards.
Governing
xAI acquired and reactivated a dormant Mississippi gas plant to supply dispatchable power directly to its Grok AI training data center — one of the first direct plant acquisitions by an AI company.
04
The Creator Is the Channel

Brands are discovering creators aren't the media buy — they're the co-manufacturer

+

The most efficient distribution channel in consumer commerce in 2026 is not a search algorithm, a retail shelf, or a paid media buy. It is a creator with an authentic relationship with an audience. Target launched two new creator programs — Club Target and Target Ambassadors — to drive influencer-led sales. Skincare brand Refy launched its new product by giving 12 creators full creative control with no prior approval required — and no guardrails. Thayers developed its new Hydrating Milky Mist in direct co-creation with TikTok creators Nina Pool and Shelby Ann Bell — the creators ideated the product, not just marketed it. YouTube Shopping launched AI-powered tools to match brands with the most authentically relevant creators by analyzing billions of data points. The pattern is consistent across beauty, food, fitness, and automotive: brands are not using creators as distribution. They are using creators as product development, brand truth-telling, and audience co-ownership.

⚡ The Now

The creator economy is undergoing a structural shift from 'influencer as media buy' to 'creator as co-manufacturer.' Brands that understand this shortcut years of market research and brand-building by embedding themselves inside creator communities at the ideation stage. The Refy approach — no approvals, full creative control — is not recklessness. It is a deliberate bet that authentic creator voice generates more ROI than brand-controlled messaging.

→ What's Next

The next strategic advantage isn't having the best product — it's having the most embedded community of co-creators. Brands that don't make this shift will keep spending on creator partnerships that perform like ads: expensive, forgettable, and increasingly ignored by audiences that have learned to recognize the format. The AI-powered creator matching tools from YouTube are the infrastructure layer that scales this: within two years, the ability to identify the right creator for the right product at the right moment will be automated, shifting the competitive advantage from 'who you know in the creator economy' to 'how well you can co-create.'

Chain Store Age
Target introduced Club Target and Target Ambassadors — two new programs designed to deepen engagement with social media influencers and drive sales through creator-led content.
Glossy
Refy compensated 12 creators to post content without prior brand approval for its Skin Base Skin Tint launch — a deliberate bet on authentic creator voice over brand-controlled messaging.
Glossy
Thayers co-developed its new Hydrating Milky Mist with TikTok creators Nina Pool and Shelby Ann Bell — the creators ideated the product concept, not just marketed it.
Glossy / YouTube
YouTube Shopping launched AI-powered tools to match brands with the most relevant creators by analyzing billions of data points — automating the creator-brand matching that has been a relationship-driven process.
YouTube
Tabitha Brown's YouTube series Tab Time won three Emmy awards — marking a milestone in the recognition of creator-originated programming as premium entertainment, not user-generated content.
05
Bodies on the Balance Sheet

The addressable market just expanded from sick patients to everyone with a body

+

The largest pharmaceutical and biotech capital allocation in history is converging on a single thesis: the human body is becoming an addressable market for precision intervention at scale. BioNTech, having collapsed from €19B revenue to €118M per quarter post-COVID, is deploying its €16.8B cash reserve into 15 Phase 3 oncology trials through 2026. Lilly announced an additional $4.5B investment in its Indiana manufacturing complex for obesity and genetic medicine. Regeneron received the first-ever FDA approval for a neurosensory gene therapy. AstraZeneca's AI framework Reinvent has halved the time to identify molecular structures for new medicines. Recursion Pharmaceuticals is building an AI model to optimize clinical trial patient recruitment. eClinical Solutions published data showing 241% ROI from AI-powered clinical trial platforms. The compounding effect: AI is accelerating discovery, gene therapy is arriving faster than regulatory frameworks can absorb it, and GLP-1 drugs have opened mass-market appetite for pharmaceutical body optimization that is expanding the population from 'sick patients' to 'everyone with a body.'

⚡ The Now

The pharmaceutical industry's growth trajectory is beginning to resemble tech's 2010s more than its historical norms. The combination of AI-accelerated discovery timelines, gene therapy at commercial scale, and mass-market appetite for biological optimization is creating a market measured in trillions — not billions. The picks-and-shovels play is the discovery infrastructure: AI drug frameworks, clinical trial platforms, gene delivery systems, and manufacturing capacity.

→ What's Next

The bet is not just better drugs — it's that the combination of precision targeting, faster development, and expanded indication will redefine what a pharmaceutical company is. BioNTech's move is the most instructive: a company that became briefly the most valuable in the world on a single product is now deploying that windfall into a platform bet on oncology. The companies building discovery infrastructure in 2026 will be operating at a structural advantage when the next generation of therapies reaches commercial scale in 2030–2035.

Drug Discovery & Development
BioNTech is refocusing on oncology with 15 Phase 3 clinical trials planned through 2026, deploying its €16.8B cash reserve into what it calls its original mission.
Fierce Pharma
Eli Lilly announced a $4.5B investment to expand its Lebanon, Indiana facilities for GLP-1 obesity drug manufacturing and genetic medicine production.
Genetic Engineering & Biotechnology News
Regeneron received accelerated FDA approval for Otarmeni — the first gene therapy to treat severe sensorineural hearing loss, marking a milestone for neurosensory gene medicine.
Fierce Biotech
AstraZeneca's open-source AI framework Reinvent has halved the time required to identify molecular structures for new medicines — now used broadly across the industry.
Fierce Biotech
Recursion Pharmaceuticals is developing an in-house AI model to optimize patient recruitment for clinical trials, analyzing a database of 3M+ patient records to identify optimal candidates.
Drug Discovery & Development
eClinical Solutions published a third-party study showing 241% ROI over three years from its AI-powered elluminate clinical trial data platform, with significant operational efficiency gains.
06
The Reshoring Paradox

Tariffs failed to move factories home — and accidentally triggered something more interesting

+

America's bet that tariffs would reshore manufacturing is failing on its own stated terms — and revealing something more consequential in the process. Kearney's 2026 Reshoring Index shows manufactured goods imports are still rising despite the highest tariff regime in a century. The promised factory jobs are not materializing at scale. But a different kind of reshoring is happening: companies are investing in US-based automation and AI-powered manufacturing to make domestic production economically viable at higher cost structures. Ford is absorbing higher per-vehicle costs rather than abandoning domestic production. Apple is reinvesting tariff refunds into US manufacturing and innovation. The US-Taiwan chip partnership is deepening, targeting 40% of domestic advanced chip demand from US fabs by decade-end. Morgan Stanley analysis shows tariffs intended to reshore steel are having limited domestic manufacturing effect — but are accelerating automation investment in response to higher input costs. The paradox: tariffs are failing to move production home at scale, but they are successfully accelerating the automation investment that makes domestic production defensible.

⚡ The Now

The real reshoring dividend is not factory jobs — it's factory intelligence. Companies using this forced-march moment to build AI-augmented, highly automated domestic manufacturing capabilities will have cost structures and supply chain resilience that post-tariff global competition can't easily undercut. The capital allocation decision that pure market economics would have deferred for another decade is being made now, under duress.

→ What's Next

The factories that matter in 2035 are being built in 2026–2028. Not because they moved jobs home, but because they built the manufacturing intelligence that defines the next era. Automation justification is already shifting — from 'labor displacement' to 'capacity augmentation.' The companies that internalize this reframe will invest faster and more strategically than those still debating whether reshoring is 'real.' The geopolitical disruption is the forcing function. The automation is the prize.

Material Handling & Logistics
Kearney's 2026 Reshoring Index finds manufactured goods imports still rising despite record US tariff pressure and reshoring investment — the stated policy goal is not being achieved.
Plant Engineering
Manufacturers are shifting automation justification from labor displacement to capacity augmentation amid supply chain disruptions and reshoring pressures — changing how automation capex is evaluated.
Stimson Center
The Trump administration is pursuing onshoring of chip manufacturing with a target of fulfilling 40% of domestic advanced chip demand from US fabs by end of decade, deepening the US-Taiwan partnership.
Morgan Stanley
Morgan Stanley economists found that while tariffs intended to encourage reshoring of steel production show limited domestic manufacturing effect, they are accelerating automation investment to offset higher input costs.
Ford Media Center
Ford reaffirmed its commitment to domestic production and implementing a 'Skunk Works' model for autonomous team innovation — absorbing higher per-vehicle costs as a strategic choice, not a constraint.
07
The Humanoid Is Early (But Not Wrong)

Physical AI is arriving faster than enterprise can absorb it — and the deployment gap is real

+

Figure's Helix-02 demo landed this week: two humanoid robots collaboratively resetting a bedroom in under two minutes using a single learned Vision-Language-Action policy — no task-specific programming. Oxford Robotics Institute partnered with AtkinsRéalis to deploy physical AI robots specifically in nuclear environments, where human access is limited and precision is existential. Accenture Ventures invested in General Robotics and its GRID platform, which unifies AI intelligence across robots from multiple manufacturers into a single operational layer. DARPA is funding Boston Dynamics for two military robot programs: ATLAS (humanoid, human-like locomotion and manipulation) and CHEETAH (quadruped, high-speed). The AI hardware investment surge that followed the software surge is arriving in physical form. But the deployment reality is stark: most humanoid applications are still controlled demonstrations. The enterprise is not ready for autonomous physical agents in unstructured environments at scale.

⚡ The Now

The Helix-02 bedroom demo is the equivalent of early LLM demos that could write a paragraph but couldn't maintain coherent long-form context. The capability trajectory is clear. The deployment gap is equally real. The companies that will win in physical AI are not the ones with the most impressive demo — they are the ones that identify the highest-value, lowest-variability physical tasks (nuclear inspection, warehouse picking, hazardous site remediation) and build reliable deployment stacks for those specific domains first.

→ What's Next

The humanoid is early. It is not wrong. The question is not 'will robots be useful?' — it's 'which physical environments will be reshaped first, and by whom?' Nuclear sites, warehouses, and hazardous industrial environments are the beachheads. The general-purpose humanoid in unstructured consumer environments is five to eight years away. The enterprise-specific physical AI in constrained, defined domains is arriving in 2027–2029. The infrastructure investment decisions being made now — platforms, training environments, governance layers — will determine who owns the physical AI stack when it matures.

Figure
Figure demonstrated two Helix-02 humanoid robots collaboratively resetting a bedroom in under two minutes using a single learned Vision-Language-Action policy — the most capable multi-robot physical AI demo to date.
Engineering and Technology
Oxford Robotics Institute and AtkinsRéalis formed a partnership to translate advanced physical AI research into autonomous systems for nuclear environments — one of the highest-stakes physical AI deployments.
Automation World
Accenture Ventures invested in General Robotics' GRID platform, which unifies AI intelligence across robots from multiple manufacturers — positioning Accenture to offer cross-fleet physical AI integration to enterprise clients.
Naval Technology / DARPA
DARPA is funding Boston Dynamics to develop ATLAS, a humanoid robot mimicking human locomotion and manipulation, and CHEETAH, a high-speed quadruped — advancing military physical AI.
IEEE Spectrum
Researchers developed a neural network using reinforcement learning and multi-expert distillation that gives robot hands human-like dexterity for complex manipulation tasks.
IEEE Spectrum
A broad industry survey finds humanoid robot manufacturers are scaling production but deployment use cases remain narrow and constrained — the gap between capability and enterprise deployment is real.

Frontier Science Feeding the Machine

The research signals underneath the enterprise stories. These breakthroughs — in gene therapy delivery, materials science, quantum mechanics, AI architecture, and climate modeling — are the scientific substrate feeding the next wave of commercial disruption.

Gene Therapy
Researchers delivered a protective gene (APOE3 Christchurch) to mouse livers — not brains — and cleared amyloid from brain tissue. The liver manufactured and circulated the therapeutic protein to the brain, bypassing the blood-brain barrier. If it translates to humans, one of the hardest delivery problems in neurology just got dramatically more tractable.
Materials Science
Seoul National University researchers developed a mesoscale carbon fiber lattice using a 3D node winding method that achieves aluminum-level mechanical performance at one-hundredth the weight. The manufacturing method is scalable — with implications across aerospace, automotive, construction, and wearable technology.
Soft Robotics
Harvard researchers developed a 3D printing technique that deposits active liquid crystal elastomers alongside passive ones, creating programmable artificial muscles that bend, twist, and contract on demand. This is not incremental — it's a fabrication method that changes what soft robotics and medical devices can do.
Quantum Computing
Harvard SEAS demonstrated for the first time that a single quantum of vibrational energy (a phonon) can interact strongly with a single electron spin — enabling spin-phonon coupling as a new mechanism for quantum communication. This advances quantum networking by providing a new, more controllable interface between quantum memory and quantum transmission.
Clinical AI
Researchers launched OncoAgent, an open-source clinical AI system for oncology using a dual-tier LLM architecture with Corrective RAG — designed to deliver multi-agent clinical reasoning while keeping patient data on-premise. The first practical multi-agent AI for cancer care that doesn't require data to leave the institution.
IoT / Supply Chain
Wiliot's Gen3 IoT Pixel is a battery-free sensing device the size of a postage stamp that connects any physical product to the cloud using ambient RF energy. At mass scale, supply chains can track temperature, humidity, and handling conditions across every item in real time — without batteries, charging, or manual scanning.
AI Architecture
Together AI announced DeepSeek-V4, a model architecture supporting 1 million token context via a hybrid attention mechanism that breaks the quadratic scaling constraint at long contexts. Million-token context is now economically viable for production inference — changing what agents can hold in working memory mid-task.
Biotech
ParcelBio's APEXm platform uses the cell's native RNA-stabilizing machinery to dramatically extend mRNA stability and therapeutic duration. Current mRNA drugs degrade quickly and require cold chain logistics — APEXm could enable mRNA medicines that persist longer and travel without extreme refrigeration.
Space Intelligence
Planet Labs and Google are collaborating on Suncatcher — a satellite constellation designed to process imagery off-planet, transmitting only analyzed insights rather than raw data. The bandwidth savings make near-real-time global monitoring commercially viable at scale for the first time.
Climate Science
Harvard SEAS researchers proposed a new explanation for the 56-million-year Sturtian glaciation ('Snowball Earth') using a coupled climate-carbon cycle model revealing a specific ocean-atmosphere tipping point. The finding has direct relevance for understanding modern carbon cycle stability thresholds and feedback loop sensitivity.