AI Funding Wave Hits Sales, Markets, Video, and Defence

Key Points

• Five AI startups raised a combined $1.4B+ across Series A, B, and D rounds in mid-2026.
• Rounds span B2B sales, financial markets, video intelligence, defence autonomy, and chip architecture — all vertical-specific AI bets.
• Quantum Systems' $1.2B Series D alone values the German drone maker at approximately $8 billion, more than doubling its valuation from under a year ago.

Sources

Key Points

• Investors are explicitly rewarding AI companies that own proprietary contextual data inaccessible to general-purpose LLMs, not just those using AI as a feature layer.
• The shift from AI-assisted tools to fully agentic execution — where AI agents act autonomously within live workflows — is accelerating across every vertical represented here.
• Defence AI is now attracting institutional capital at scale: Blackstone's participation in Quantum Systems signals that mainstream asset managers are entering a sector previously dominated by specialist funds.
• The Amazon-TwelveLabs cloud partnership illustrates how large hyperscalers are using venture co-investment to lock in long-term infrastructure customers, blurring the line between strategic partnership and financial investment.
• OXMIQ's IP-licensing model for GPU architecture challenges NVIDIA's dominance by lowering the barrier for custom AI silicon without requiring full chip fabrication — a capital-efficiency play with asymmetric upside.

Sources

Sources

winners

• Enterprise sales teams using Aligned: 30% faster cycles and 15% higher win rates are quantified, defensible advantages over competitors still running deals via email chains.
• Institutional investors on LinqAlpha: context-driven AI that compounds on proprietary research history creates alpha differentiation that generic tools cannot match.
• AWS and Amazon: the TwelveLabs partnership converts a $100M investment into a long-term infrastructure anchor tenant, with models optimised for Trainium chips.
• European defence primes and NATO allies: Quantum Systems' expanded production capacity and MOSAIC platform directly increase the available supply of proven autonomous systems.

losers

• Legacy CRM and revenue intelligence vendors: Aligned's 'system of action' framing directly positions Salesforce, HubSpot, and Gong as incomplete background infrastructure rather than deal-critical software.
• Generic financial AI platforms: LinqAlpha's firm-specific training model makes one-size-fits-all research tools structurally less valuable for institutional clients with deep research libraries.
• NVIDIA's GPU moat: OXMIQ's open, licensable GPU architecture — with CUDA compatibility baked in — directly targets the switching costs that underpin NVIDIA's enterprise lock-in.
• Traditional video asset management vendors: TwelveLabs makes static video search tools obsolete for any organisation with large unstructured footage archives.

implications

• The 'vertical AI' investment thesis is now confirmed at scale: generalist AI is table stakes; domain-specific agents with proprietary data loops are the defensible business.
• Cloud hyperscalers are evolving their venture strategies from passive capital allocation to active infrastructure pre-commitment, with TwelveLabs-AWS as the clearest example yet.
• European defence tech has cleared a credibility threshold with institutional capital: the $8.7B raised by European defence startups in 2025 — cited in the Quantum Systems round — is not a blip but a structural reallocation.
• The GPU supply chain is entering a phase of architectural pluralism: OXMIQ, alongside Tenstorrent and others, signals that the era of a single dominant GPU architecture is under genuine competitive pressure.

minority report

• Aligned's 'system of action' narrative may be overstated: Salesforce, Microsoft, and HubSpot all have the distribution, data integrations, and AI budgets to replicate a buyer-facing deal workspace within 18 months, potentially commoditising the category before Aligned achieves sufficient enterprise lock-in.
• LinqAlpha's firm-specific training model assumes institutional investors will freely share proprietary research history with a third-party vendor — a compliance and IP risk that may slow adoption at the largest, most regulated asset managers.
• Quantum Systems' $8B valuation at profitability is striking, but the drone market has historically seen rapid commoditisation once hardware specs become standardised; software differentiation via MOSAIC must be proven at scale before the valuation is structurally justified.

Sources

second order

• As agentic AI embeds in sales workflows, the role of human sales representatives will bifurcate: top-performing reps will leverage agents to dramatically increase deal capacity, while mid-tier reps face meaningful displacement pressure.
• TwelveLabs' AWS optimisation for Trainium chips creates a precedent: future AI model companies may negotiate cloud exclusivity arrangements as part of their venture rounds, reshaping standard term sheet structures.
• Quantum Systems' profitability at $8B valuation, combined with Blackstone's entry, is likely to trigger a revaluation of other European defence AI companies currently trading at discounts to their US counterparts.
• OXMIQ's CUDA-compatible licensing model, if adopted by even one major device manufacturer, would create a parallel GPU ecosystem that forces NVIDIA to compete on architectural merit rather than ecosystem lock-in alone.

prediction

• Within 12 months, at least one major CRM vendor — most likely Salesforce or HubSpot — will acquire or launch a direct competitor to Aligned's AI Deal Workspace, validating the category while accelerating the competitive clock.
• Quantum Systems will file for an IPO or pursue a SPAC-equivalent listing on a European exchange by late 2027, targeting a valuation above $12B on the back of expanded NATO contract wins.
• TwelveLabs will announce a second major cloud partnership — likely with Microsoft Azure — within 18 months, as hyperscalers race to secure video intelligence capabilities for enterprise customers.
• OXMIQ will close its first commercial chip architecture license with a semiconductor or robotics OEM within 12 months, proving the IP-licensing model and triggering a Series B at a significantly higher valuation.

minority report

• The entire 'vertical AI agent' investment wave may be timing-sensitive in a way investors are underpricing: if frontier model capabilities advance rapidly enough in 2026-2027, the proprietary data moats these companies are building may be neutralised by models that can generalise across domains without firm-specific fine-tuning — compressing the window of defensibility significantly.
• Quantum Systems' rapid valuation doubling may reflect geopolitical risk premium rather than durable business fundamentals: a ceasefire in Ukraine or a reduction in NATO spending commitments would materially impact order books and make the $8B valuation difficult to sustain.
• OXMIQ's open GPU licensing model, while intellectually compelling, faces a distribution problem: chip architecture licenses require deep integration partnerships with foundries and OEMs that typically take years to negotiate, meaning the capital-efficient model may prove slower to revenue than the Series A narrative implies.

What This Means

Detected Trends

Sources