● BIOLOGY IS CODE — THE BIOLOGICAL OPERATING SYSTEM
We are moving from reactive chemistry to a predictive Biological OS
Companies bridging the gap between molecular reality and AI computation are generating the largest asymmetric wealth-creation event of the decade. The thesis reduces to three verbs: Read the code, orchestrate the data, write the biology.
Read the CodeOrchestrate the DataWrite the Biology
Snapshot · Aug 17, 2026 · 9 tickers
The new scaling law is Healthspan per Token
Old scaling laws governed software: more parameters → more intelligence. The next massive consumer reallocation of capital is biological. Consumers exhibit near-full price elasticity for platforms that demonstrably add years of healthy life — bending the curve from traditional aging (1 year lived = 1 year lost) toward longevity escape velocity, where predictive ontologies return more than a year of healthspan per year lived.
"SICKNESS IS SIMPLY A BUG IN THE PROTEOMIC NETWORK"
Level 5Clinical outcomes
Human health, lifespan, systemic performance.
Level 4Proteins & proteoforms
3D Lego-like structures that execute function. Disfunction here is the root of all disease.
Level 3Peptides
Short chains — the body’s biological instruction set.
Level 2Amino acids & chemistry
Raw dietary inputs and chemical compilers.
Level 1Atoms
The lowest level of abstraction — the physical hardware.
To fix a bug at Level 5, AI must compute the physics of Level 1 and write code via Level 3. Read the proteome, then write back with precision.
MAPPING THE BIO-OS VALUE CHAIN
READ
Data extraction
Nautilus ($NAUT)
10B-protein mapping; Tau proteoforms / neuro.
Tempus & Caris ($TEM / $CAI)
Deep oncology data integrated with hospitals.
Recursion ($RXRX)
Wetlab simulation lowering chemical toxicity.
ORCHESTRATE
Distribution & inference
Hims ($HIMS)
D2C infrastructure, longitudinal biomarker aggregation — the incentive provider that owns the consumer. AI "load bearing, not decorative": the only closed loop of intake → treatment → follow-up → outcome at ~3M-subscriber scale, aiming to sell $50–150K/yr concierge-level care by subscription.
WRITE
Biological intervention
ImmunityBio ($IBRX)
Immune-system reboot (IL-15 superagonist).
MiNK ($INKT)
Immune bypass — iNKT cells / stable lipids.
NervGen ($NGEN)
Nervous-system regeneration (NVG-291).
Parabilis ($PBLS)
Solving flat-protein errors (Helicon peptides).
Read and Write are brilliant tools, but they lack the end-consumer relationship. The platform that owns distribution and provides incentives (the D2C ontological flywheel) captures the majority of the economic value — an insurmountable "singularity scaler."
MARKET CAP · Aug 17, 2026
TEM
$51
READ
Tempus AI
Deep oncology data, integrated with hospitals. Q2 reported Jul 30.
$9.3B
IBRX
$7.26
WRITE
ImmunityBio
Immune system reboot (IL-15 superagonist). Q2: continued ANKTIVA growth.
$7.7B
HIMS
$32
ORCHESTRATE
Hims & Hers Health
D2C rails, the incentive provider. Q2 rev $753M (+38%), ARPU $92 (+21%); Q3 guided to +47–50%, FY26 raised to $3.1–3.3B.
$7.3B
CAI
$19
READ
Caris Life Sciences
Molecular profiling. Record Q2 ($264M rev, +45%); turned adj-EBITDA positive.
$5.5B
PBLS
$37
WRITE
Parabilis Medicines
Helicon peptides for flat / undruggable proteins. IPO Jun 2026.
FY2022–FY2025 actuals shown. Q2 FY2026 (reported Aug 10): revenue $753M (+38% YoY), ~2.9M subscribers (+19%, +300K net adds), and monthly revenue per average subscriber of $92 vs $76 a year ago (+21%) — both curves rising while marketing fell 5 pts YoY to 34% of revenue. International revenue grew ~17x to $131M (incl. ~$40M from Eucalyptus, closed June; organic +13% QoQ), with the UK, Australia and Germany each above a $100M annualized run rate. Adjusted EBITDA $60M (8% margin, +1 pt QoQ); GAAP net loss -$86M, hit by ~$81M of one-time costs (Eucalyptus close, restructuring, $47.5M FTC accrual); gross margin compressed to 64% on branded-GLP-1 and international mix. Guidance implies acceleration: Q3 at $880–900M (+47–50% YoY), FY2026 raised to $3.1–3.3B (+32–41%) with $275–325M adj. EBITDA; 2030 targets of $6.5B+ revenue / $1.3B+ adj. EBITDA reiterated. The AI-native Hers Weight Loss rollout (July) showed 3x messaging engagement, 80% of questions answered by AI within clinical protocols, ~50% fewer non-clinical support tasks, and lower cancellations in the new-experience cohorts — management expects AI investment to pay back in 12–18 months.
FCF POR ACCIÓN · quarterly · last 6 quarters
Hims & Hers Health
HIMS · 231M sh · 6Q
Official Q2’26 deck reconciliation. Q2’26 FCF of –$68M reflects working capital absorbed by the branded-Wegovy ramp — covered with a $400M receivables facility plus a ~$400M convertible, ending the quarter with $840M+ in cash. Management expects a return to FCF generation in H2 2026.
Tempus AI
TEM · 180M sh · 6Q
Lumpy but improving: Q2’26 burn of –$18.6M is the smallest outflow in the series; Q2’25 was FCF-positive.
Volatile burn despite 30% opex cuts — partnership inflows make quarters lumpy. Cash $665M, runway to early 2028.
Nautilus Biotechnology
NAUT · 127M sh · 6Q
Metronomic ~$13M/q burn, pre-revenue. FY25 burn actually declined 15% vs FY24 — disciplined for a platform builder.
MiNK Therapeutics
INKT · 5.0M sh · 6Q
Smallest absolute burn of the basket (~$1–2M/q), but the ~5M-share float makes each quarter ≈ –$0.30/sh. Q2’26 reports ~Aug 14.
NervGen Pharma
NGEN · 81M sh · 5Q
5 quarters — its entire public quarterly history (Nasdaq Jan 2026). Burn accelerating as NVG-291 trials scale; $60M raised May 2026.
CAI
–$538M
FY25 net loss (distorted by IPO stock comp); net margin TTM already +3.8%, EBITDA TTM +$129M. IPO Jun 2025: first 6-quarter FCF series completes with Q3’26 filing.
PBLS
–$153M
Net loss, 12 months to Mar 2026 (S-1). IPO Jun 2026 with $670M raised — first public quarterly cash flow statement lands with the Q2’26 10-Q.
Discrete quarterly FCF (operating cash flow − capex) from 10-K/10-Q filings via Macrotrends, YTD converted to standalone quarters. NGEN shows 5 quarters — its entire public history. CAI and PBLS IPO'd too recently for any public quarterly series; their real annual/TTM burn is shown above instead of an invented one.
The compute required for biology dwarfs today\u2019s AI infrastructure
Every H100 GPU installed on Earth would provide only ~0.02% of the compute needed to build an exhaustive proteomic map of a single human for a single year. We are at the absolute infancy of biological computation — the singularity scalers that solve this bottleneck can eclipse the scale of today\u2019s tech giants.
METODOLOGÍA & NOTAS
Market caps and prices are a point-in-time snapshot (Aug 17, 2026) and move daily.
Income-statement charts are annual GAAP actuals (FY2022–FY2025) from company filings and Yahoo Finance. NAUT and INKT are pre-revenue, so only operating income is plotted.
Q2 FY2026 results (HIMS, TEM, CAI, IBRX, RXRX, NAUT reported late Jul–mid Aug 2026) are reflected in each ticker's commentary rather than the annual bars, since FY2026 is not yet complete. HIMS figures are from its Aug 10 Q2 deck and call.
CAI (IPO Jun 2025) shows only FY2024–FY2025; its gross profit is estimated from margin and its operating income is approximate (2025 distorted by IPO stock comp).
IBRX operating income for 2022 and 2025 is approximate (derived from R&D + SG&A).
PBLS (IPO Jun 2026) and NGEN (Nasdaq Jan 2026) are pre-revenue with limited public history — shown as info cards, not charts.
FCF per share uses discrete quarterly FCF (OCF − capex) from company filings via Macrotrends — except HIMS, whose series comes directly from the official Q2 FY2026 investor deck reconciliation — divided by current diluted shares held constant across the series to isolate the cash trend. Pre-revenue names show TTM/annual burn because no 6-quarter public series exists.
This is data and research context, not investment advice.