CB Insights

AI’s talent war is shifting. And… the shift reveals what will matter most for AI startups in 2026. The latest insights on hiring patterns across open roles at the top 0.1% of AI startups (900+ Mosaic) show that GTM roles represent 44.3% of all hiring, surpassing technical roles at 38.7%. Let that sink in. The companies building the most advanced AI technology in the world are now hiring MORE for go-to-market than for engineering. The "build it and they will come" era is over. Even with breakthrough AI capabilities, companies can't scale without enterprise sales muscle, customer success infrastructure, and partner ecosystems. So, where are AI leaders focusing for 2026? 1. Agents! Agents! Agents! 2. Enterprise! Enterprise! Enterprise! 3. Infrastructure! Infrastructure! Infrastructure! 4. Geographic Expansion 5. Security and Compliance 2026 will be the year that AI startups start to be judged on $$$ rather than just promise. The winners won't just be the companies with the best tech, they'll be the ones who can actually sell it, grow it, and retain it.

People ask me if data centers are a bubble. For now, the overcapacity is in the parts AI doesn’t use. Everything that matters to AI is still scarce. That scarcity is why value is concentrating around 7 strategic control layers. The companies that own multiple layers will dictate AI infrastructure economics for the next decade. Multi-gigawatt buildouts require city-scale electricity, but grid queues can stretch for years. Power is just the first constraint. It cascades into cooling, compute, networking, storage, cloud services, and facilities. Using CB Insights predictive intelligence, we mapped 367 companies across this $7T market to show where the advantages are concentrating. Here are the signals that emerged: → Power generation: Without reliable power, nothing else matters. Products that help developers bypass multi-year grid delays are scaling fast. → AI computing hardware: Custom silicon is a moat. It gives companies more control over performance, cost, and supply. → Supporting hardware infrastructure: AI chips generate 3–5x the heat density of traditional servers. Cooling determines deployment speed. → Networking & connectivity: Training is going multi-site because no single location has enough GPUs or power. Inference is going to the edge to cut latency and bandwidth costs. Networking becomes the coordination layer. → Data storage & database infrastructure: Vector databases still matter for embeddings, but keeping models fresh cost-effectively depends more on how fast data moves between storage and compute — no matter which system you use. → AI cloud services: Specialized players help companies get scarce GPUs, including through spot markets where you can buy capacity on short notice. They also help companies run and secure AI across different setups. This layer sets the pace of broader enterprise adoption. → Data center facilities: Owning facilities = controlling timelines. The difference between 18-month and 36-month buildouts is billions in missed opportunity. Vertical integration across 3+ layers is becoming the only way to guarantee capacity, compress timelines, and control costs. The companies that own multiple layers will set pricing. Those dependent on one layer will get squeezed. Which hyperscaler is moving fastest to lock up the full stack? The full value chain and analysis are here: https://lnkd.in/eBtb4Gcf

Salesforce's M&A playbook for dominance in the AI era: 39% of Salesforce's recent acquisitions came from companies they'd already invested in. This isn't traditional M&A. It's a multi-year predictive intelligence operation where Salesforce's 1,100+ portfolio companies are prepped and evaluated for acquisition. Now, Salesforce is applying this playbook to AI M&A. They're the most active acquirer of AI companies in 2025, the overall most active acquirer of AI agent startups, and their current corporate venture profile is filled with their next acquisitions. Their AI acquisition pipeline today: → 147 active AI investments → 37 AI agent companies backed → 46% of portfolio = top-decile startups based on CB Insights Mosaic score -- our measure of a company's probability for success → 31% showing above-average M&A probability They aren't guessing which technologies will win. They are funding strategic categories and running live experiments through their customer base. The playbook: ✓ Invest early (median 3.3 years before acquisition) ✓ Validate product-market fit through customer deployments ✓ Test integration points via partnerships ✓ Observe leadership under pressure ✓ Build technical due diligence Some acquirers buy blind. By the time Salesforce writes a check, they've already watched companies operate for 3+ years. While competitors debate "build vs. buy," Salesforce already has 147 AI companies in evaluation mode. In looking at the data, it's clear they're not catching up to the AI wave – they've been proactively positioning for it for years with their investment and M&A pipeline. Agentforce didn't appear overnight. The future of Salesforce AI is already being built, backed by 37 agent-specific investments, 147 AI bets, and years of systematic portfolio construction.

The metaverse has vanished from Meta's strategy map. What's replaced it? Personal, physical AI… artificial intelligence that lives in the real world through the most human modes of interaction – vision (smart glasses) and audio (voice AI). The evidence is clear in their latest deals and partnerships: • Smart glasses are the new metaverse: Partnerships with EssilorLuxottica, Prada, Oakley, and Ray-Ban. Ray-Ban Meta glasses already have product-market fit. This is no longer R&D, this is the first implementation of consumer physical AI at scale. • AI infrastructure at unprecedented scale: Multi-gigawatt compute clusters, partnerships spanning Google Cloud to custom ARM chips, and massive data center investments. They’re all in on the building blocks for personal superintelligence. • Energy as strategic necessity: Dense renewable partnerships aren't climate-focused moves, they're operational requirements for gigawatt-scale AI. • Audio AI everywhere: Acquisitions of WaveForms AI, PlayAI, Audio Analytic signal voice as the primary AI interface. Meta's *new* strategic thesis: the AI revolution will be won through interfaces people actually use; augmenting the physical world they inhabit, not virtual worlds they'll never visit. Was the $30B+ metaverse project just the precursor to the physical AI reality?

OpenAI and Anthropic are playing tug-of-war for key relationships. Anthropic just had its best month ever, capturing its highest share yet (37.6%) of new, key business relationships. The latest data highlight the battle between OpenAI and Anthropic: 1) Anthropic is picking up steam, accelerating its share of new business relationships versus OpenAI in the last 6 months. 2) 7% of companies have key relationships with both OpenAI and Anthropic – these overlaps are concentrated among the largest enterprise software providers and big tech companies. The mid-November Microsoft-NVIDIA-Anthropic deal is exemplary, as even OpenAI’s main backer diversifies its bets. 3) Despite lower absolute revenue ($7-9B ARR vs OpenAI's $13-20B), Anthropic projects breakeven by 2028 – potentially two years ahead of OpenAI's 2030 target. 4) Anthropic is winning the most ground in enterprise software and professional services – two sectors that have outsized distribution and usage patterns that are consistent with Anthropic’s strengths. While OpenAI still has more overall relationships, the shift in momentum of new key relationships adds to the building stories around Anthropic’s relative efficiency (driven by an enterprise-first approach) and the sustainability of OpenAI’s financial model. The next six months of new business relationships will be telling. *CB Insights’ business relationships data tracks relationships like partnerships, supplier-distributor, or client-vendor to provide predictive intelligence on company strategy and competitive dynamics.

NVIDIA is building the AI economy’s operating system. The $2B investment in Synopsys Inc is a clear signal as NVIDIA’s deal pace continues to accelerate. In 2025 alone, NVIDIA added 400+ new key business relationships (up 30% vs. 2024), executed 34 direct deals (up 70%), plus 42 NVentures and 69 Inception Program investments -- alongside 12 acquisitions in the last two years. This year’s flurry of activity clusters around 4 pillars: → AI infrastructure dominance: Deals with Oracle, Digital Realty, Pure Storage, and Cisco, plus investments in Crusoe and Nscale, all reinforce NVIDIA’s role as the backbone of global AI compute. → Edge AI + robotics: Partnerships with Joby, Uber, Foxconn, and Stellantis in addition to acquisitions in model optimization and efficient edge deployment (OmniML, Deci, Brev) give NVIDIA an edge in bringing AI to the physical world. → Industry-specific applied AI: Deals with Eli Lilly, ServiceNow, and Palantir reflect a push into vertical AI models and domain-tailored copilots. → Semiconductor + networking moat: The $2B Synopsys investment --combined with Ayar Labs, Mellanox, Cumulus, and Run:ai -- expands their chip-to-cloud control plane. What's NVIDIA’s next move? The predictive intelligence signals across partnerships, investments, evolving startup markets, and competitive moves highlight sovereign AI ecosystems, enterprise AI software, chip-to-cloud networking, and sector-specific models as key target areas. P.S. Strategy Maps that show how companies are moving, what’s changing, and where momentum is building are now live on CB Insights! Our AI generates them based on our unique business relationship data.

Eli Lilly just became pharma's first $1 trillion company! Where are they investing, acquiring, and partnering for the next phase of growth? As many pharma leaders are downsizing internal R&D and making mega-acquisitions, essentially becoming manufacturing engines powered by external innovation. Lilly is taking a mixed approach. They scaled R&D spending 27% YoY to $3.47B while building a distributed innovation network of 100+ partnerships across AI drug discovery, gene therapy, RNA therapeutics, and radiopharmaceuticals. Key strategic bets across partnerships and investments: • Oral GLP-1 (orforglipron) launching 2026 • AI-first drug discovery across 13+ platform partnerships • Direct-to-consumer channels bypassing traditional insurance gatekeepers • Over $50B committed to manufacturing expansion to meet surging demand • Blood-brain barrier tech for Alzheimer's prevention trials • Pipeline diversification: 16 new Phase 3 programs since 2024 This isn't diversification for its own sake. It's a company using unprecedented cash flows to systematically de-risk future growth across multiple therapeutic frontiers. P.S. Strategy Maps that show how companies are moving, what’s changing, and where momentum is building are now live on CB Insights!

Every month, we launch new predictive intelligence capabilities that help our customers make their next move first — identifying the right companies, markets, and opportunities before competitors do. Here are the latest upgrades: ⚡𝗘𝗮𝗿𝗹𝘆 𝘄𝗮𝗿𝗻𝗶𝗻𝗴𝘀 𝗼𝗻 𝗰𝗼𝗺𝗽𝗲𝘁𝗶𝘁𝗼𝗿 𝗺𝗼𝘃𝗲𝘀: 𝗡𝗲𝘄 𝗖𝗼𝗺𝗽𝗲𝘁𝗶𝘁𝗶𝘃𝗲 𝗜𝗻𝘁𝗲𝗹 𝗔𝗴𝗲𝗻𝘁  Stop learning about competitor moves from press releases. The Competitive Intel Agent looks across changes in your competitors’ hiring, partnerships, investments, product activity, and other signals we track – then tells you what it all means for your business. You get an instant, ready-to-share threat briefing. 🗺️ 𝗦𝗲𝗲 𝘁𝗵𝗲 𝗯𝗶𝗴 𝗽𝗶𝗰𝘁𝘂𝗿𝗲 𝗯𝗲𝗵𝗶𝗻𝗱 𝗮𝗻𝘆 𝗹𝗶𝘀𝘁 𝗼𝗳 𝗰𝗼𝗺𝗽𝗮𝗻𝗶𝗲𝘀: 𝗬𝗼𝘂𝗿 𝗪𝗮𝘁𝗰𝗵𝗹𝗶𝘀𝘁𝘀 𝗷𝘂𝘀𝘁 𝗴𝗼𝘁 𝗮𝗻 𝘂𝗽𝗴𝗿𝗮𝗱𝗲 Our Watchlists help you track and evaluate companies.  Now, they include Activity Dashboards and Strategy Maps that show how companies are moving, what’s changing, and where momentum is building. Spot early shifts in investment, partnerships, and strategy – long before they’re obvious. 🚀 𝗖𝗼𝗺𝗽𝗮𝗻𝘆 𝘀𝗰𝗿𝗲𝗲𝗻𝗶𝗻𝗴 𝗮𝗻𝗱 𝗰𝗼𝗺𝗽𝗮𝗿𝗮𝘁𝗶𝘃𝗲 𝗮𝗻𝗮𝗹𝘆𝘀𝗶𝘀 𝘄𝗶𝗹𝗹 𝗻𝗲𝘃𝗲𝗿 𝗯𝗲 𝘁𝗵𝗲 𝘀𝗮𝗺𝗲: 𝗜𝗻𝘁𝗿𝗼𝗱𝘂𝗰𝗶𝗻𝗴 𝗜𝗻𝘀𝗶𝗴𝗵𝘁 𝗖𝗼𝗹𝘂𝗺𝗻𝘀 (𝗿𝗲𝗾𝘂𝗲𝘀𝘁 𝗲𝗮𝗿𝗹𝘆 𝗮𝗰𝗰𝗲𝘀𝘀) We’re not done improving Watchlists. You can now add Insight Columns, bringing our full AI-powered analysis across an entire list in one step. Ask a question once and instantly populate a column with insight for every company — all powered by CB Insights data. See how they stack up, where momentum is building, catch shifts you’d otherwise miss, and narrow your short list fast. You’ve never seen anything like it. 𝗖𝗕 𝗜𝗻𝘀𝗶𝗴𝗵𝘁𝘀 𝘅 𝗖𝗹𝗮𝘆: 𝗖𝗼𝗺𝗽𝗿𝗲𝗵𝗲𝗻𝘀𝗶𝘃𝗲 𝗖𝗥𝗠 𝗶𝗻𝘁𝗲𝗴𝗿𝗮𝘁𝗶𝗼𝗻𝘀 𝗲𝘅𝘁𝗲𝗻𝗱𝗲𝗱 𝗳𝘂𝗿𝘁𝗵𝗲𝗿 Many of the world’s best GTM teams integrate CB Insights proprietary data into their CRM, using our integrations with Salesforce and other leading platforms. With our new Clay partnership, GTM teams running on Clay have an additional option to inject CB Insights predictive intelligence into their CRM, so they can pinpoint when a prospect enters a key growth phase, engage at the perfect moment, and close more deals faster. Learn more: https://lnkd.in/gPKN6c9v

BNPL: Build Necessary Partnership Layer The next era of Klarna vs Affirm hinges on partnerships, not merchants – and only 6.3% of their current key relationships overlap. That means Klarna and Affirm are building mutually exclusive pipes into the commerce layer. In the agentic era, these partnerships become the rails that AI agents will shop on. Why partnerships will define BNPL player dominance in agentic commerce: 1) Partnership depth > breadth: AI agents won't randomly discover merchants; they'll route through deeply integrated platforms. Provider distribution through strategic relationships matters more than merchant reach. 2) The "AI agent fee" emerges: Affirm's $2,447 revenue per merchant (2.3x higher) reflects diversified monetization across merchant fees + consumer products. This multi-sided model creates flexibility to structure agent platform deals differently than Klarna's merchant-fee-dependent model. 3) Platform consolidation accelerates: The major players (and their agents) will pick preferred BNPL partners. Agents will execute within the guardrails of these partnerships. Integration depth > merchant breadth. Strategic relationships formed NOW determine who's embedded in the agent layer tomorrow. The battleground is shifting from winning merchant count to becoming the default payment rails for agentic commerce. Partnerships need to be Built Now so that service providers can Profit Later.

What is the future of Physical AI? Key takeaways from my keynote at the Plug and Play Summit yesterday: → Agentic AI and Physical AI are incremental alone, transformative together. Orchestrated AI creates multiplicative value by coordinating autonomous systems across entire value chains. → Big tech and industrial giants are battling to own the orchestration layer, with high-momentum startups forcing the pace. → Agentic AI unlocks the robotic brain. Reasoning and inference capabilities finally solve the decades-old bottleneck. → Humanoid funding will hit 3x the previous record this year. The leading companies rank in the top percentile of all startups on our platform. → Open vs closed ecosystems will define competitive moats for the next decade. → Robotics and AI engineers are the scarcest asset in the entire stack. Across the data, we are seeing that partnerships, investments, acquisitions, and internal development are all accelerating simultaneously as companies race toward AI's impact in the physical world. Thank you to Plug and Play Tech Center and the Plug and Play Mobility & Physical AI team for having me deliver CB Insights' predictive intelligence to kick off an incredible summit!