Section D · Position

Outlook

Together is one of the most interesting bets in the AI infrastructure landscape because it sits at the intersection of open-source momentum and managed-inference economics. The strategic question is which forces — hyperscaler absorption, open-source continued growth, frontier-capability gap — dominate over the next several years.

Bull case

  1. Open-source maintains capability momentum. Llama, Qwen, DeepSeek, others continue narrowing the gap with closed-source frontier; Together's catalog stays competitive.
  2. Inference market grows faster than training. The shift toward inference (driven by AI product deployment) plays to Together's strength.
  3. Cost-sensitivity rises. As AI adoption scales, the per-token cost differential matters more to more customers.
  4. Research advantage compounds. Together's optimization stack stays ahead of competitors; unit economics improve.
  5. Lifecycle bundling captures share. Customers value the inference + fine-tuning + training combo; Together captures lifetime value other platforms can't.

Bear case

  1. Hyperscalers absorb open-source. AWS / Azure / GCP make Together's value proposition redundant for customers already in their orbit.
  2. Open-source quality stalls. Frontier closed-source pulls further ahead; the open-source segment becomes a niche rather than a mainstream alternative.
  3. Margin compression. Inference becomes commoditized; per-token prices fall faster than cost-per-token; margins compress.
  4. Closed-source pricing pressure. If OpenAI / Anthropic cut prices significantly, the open-source cost advantage narrows.
  5. Specialist competitors fragment market. Fireworks, Lepton, Anyscale, others each take pieces of the inference market.

Scenarios

Scenario A: The platform leader

Together becomes the dominant managed-inference platform for open-source models. Multi-billion-dollar annual revenue; eventually a public listing. Strategic value as the AI ecosystem's open-source counterpart to closed-source frontier.

Scenario B: Profitable specialist

Together carves out a sustainable position in the open-source inference market at moderate scale. Profitable software business; maintained as private or modest exit.

Scenario C: Squeezed by hyperscalers

Hyperscaler-hosted open-source absorbs the broader market; Together is left with the customers who specifically want the platform's research-credibility and lifecycle-platform features. Smaller but persistent.

Scenario D: Acquisition

A hyperscaler or AI company acquires Together for the research talent and platform technology. Premium price; consolidation move.

What to watch

  • Open-source model release cadence and capability progression.
  • Closed-source pricing moves (OpenAI / Anthropic / Google).
  • Hyperscaler open-source-inference offering evolution.
  • Together's revenue growth disclosures or reports.
  • Major customer wins or strategic partnerships.
  • IPO signals or other liquidity events.
  • Research output continuity.

Takeaway

Together is one of the cleanest tests of the open-source-AI thesis as a commercial business. The company's success or struggles will tell us how durable the open-source-friendly slice of the AI market really is. After this guide, Lambda is the natural next read — a longer-running player in a different segment of the same landscape.