Spring is here — and so are the most significant updates shipped to the Helios AI platform. This document is the master release notes log: what's live and what's coming next.

The biggest upgrade to the Helios AI climate risk engine since launch. Three major improvements now in production:

Every crop's growing calendar has been rebuilt with sub-period precision — including perennial crops with multi-year overlapping cycles. The system now knows which exact growth stage a crop is in at any given moment and scores climate risk against that stage in real time. Example: Helios knows May marks the Fruit/Crop Fill phase for Cocoa in Peru and adjusts risk accordingly — not just by month, but by crop phase.

When climate risk spikes, Helios cross-references current conditions against a database of 500+ confirmed historical disruption events, matched side-by-side with the actual yield outcomes that followed. The system now surfaces comparisons like: "The last time conditions looked like this was 2022 — and yields dropped 8%." When a confirmed disruption match is detected, the platform links directly to the original source article — so teams can verify and act, not just react.

Every alert is now backed by crop-stage awareness, compound risk modeling, and historical event matching working together as a unified engine — not as independent signals.

Real-time risk signals delivered directly to your inbox. Each morning, if risk has shifted across any commodity and country you track, you receive a digest grouped by commodity — showing exactly what changed, where, and by how much.

Powered by the full Climate Risk 2.0 engine: crop-stage awareness + compound risk modeling + historical event matching.

Configure: Account Settings → Email Notifications

Two new commodities released to the platform:

Major releases coming to the platform. Ordered below from broadest scope (cross-platform synthesis) to deepest layer (data reliability):

The biggest new concept in this release cycle. Helios AI has always told you what the climate is doing — the Disruption Impact Center tells you what that means for supply. Every active disruption is tracked from early signal to confirmed market event, with quantified yield and supply impact at every stage.

Each event moves through a defined lifecycle — Emerging → Predicted → Active → Confirmed (with news coverage badge) — so you always know exactly where a disruption stands, not just that one exists.

Trigger → Yield Impact → Supply Impact. When drought hits Côte d'Ivoire during cocoa's critical flowering stage, you see the estimated yield decline (e.g., 12–18%) and the resulting global supply reduction (~6%) — not just a risk score.

Quantifies how much of a yield change is climate-driven vs. other factors. When a supplier cites weather-related cost pressures to justify a price increase, you have independent data to verify or push back in the same conversation.

Bumper conditions follow their own lifecycle (Favorable → Surplus Forecast → Bumper) — so supply surpluses driving prices down surface alongside risk events on the same dashboard.

Real-time lifecycle transition alerts + Monday morning digest. Zero configuration required; if your tracked commodities aren't affected, your inbox stays quiet.

Climate, Supply & Yield, and Cash Prices pages now surface disruption banners when an active or predicted event is relevant to the commodity you're viewing.

KPI strip (Active / Predicted / Bumper-Surplus / Emerging event counts), Most Impacted Commodities panel, and a What Changed Today feed showing every lifecycle transition since your last visit — all scoped to your portfolio.

Previewed at the April demo webinar. This is the full release.

The Supply & Yield page is getting a significant depth upgrade: 10 years of historical data, 3-year forward forecasts, and a new Yield tab — across 57 commodities and 86 countries.

Available across every view — global and country-level, production and yield. A year of context tells you what's happening now. A decade of context tells you whether it's normal.

Every forward forecast includes a range that widens over time — an honest representation of how uncertainty compounds. Alongside this, historical accuracy indicators show how closely models have tracked actual outcomes over the past 3 years.

Yield (metric tons per hectare) is a different question than production volume. When production drops, yield data tells you whether farmers are harvesting less land or getting less from the land they're harvesting — and those two situations have very different implications for how durable the supply trend is. Global and country-level yield forecasts now available across all covered commodities, with sub-national (state/regional) breakdowns for major producers. Includes Helios Horizon AI-generated explanations synthesizing weather patterns, seasonal conditions, and historical context into plain-language analysis.

Production and area harvested, side by side:

On the Supply tab, both metrics are now available at the same depth — because production tells you the output, area harvested tells you the input, and yield is what connects them. Every country view includes comparison to its 10-year historical average.

At the country level, full distribution of likely outcomes for the upcoming season — not just the midpoint, but the range of scenarios and how probable each one is. Tight distribution = predictable season. Wide distribution = meaningful uncertainty to plan around.

For any country, brings together beginning stocks, imports, production volume, exports, domestic consumption, and ending stocks — historical and forecast — in one place. Production is only one input into net availability.

Tighter climate-to-supply integration (when climate shifts in a major producing region, downstream production implications surface automatically); expanded yield prediction coverage; richer market data connecting supply intelligence to the pricing layer.

Understanding a crop's risk profile just got significantly faster. Helios now surfaces an AI-generated narrative explanation at the top of every country and regional crop page — telling you not just what the current rating is, but why.

Example output:

India's lentil crop faces severe climate stress during critical yield-formation periods, threatening significant production losses nationwide. Major producing states — Madhya Pradesh (39%), Uttar Pradesh (32%), Bihar (8%) — are all experiencing above-normal heat and moisture deficits at the worst possible time in the season.

Available at both country and regional level. Fully responsive across desktop, tablet, and mobile.

The accuracy bar was previewed at the April demo webinar. This is the full, formal release details.

Two new capabilities now on every price chart across every commodity, price type, and geography on the platform.

Every price forecast is saved at generation time. The platform now plots historical accuracy dots alongside the actual price line — showing what Helios predicted the price would be one month before any given date, and how that prediction compared to reality. Above the chart, a 1-Month Historical Accuracy bar summarizes performance as a percentage, specific to that exact combination of commodity, origin, variety, and price type. The bar gives you the summary; the dots let you see the full picture underneath.

On the forecasted side of the chart, a price range band now surrounds every forward prediction. Narrower near the current date, widening as you move out — because more can change over a longer horizon. The range accounts for model performance on that specific series, data quality and depth, and the natural increase in uncertainty over time. This replaces historical price ranges (which described where prices had been) with ranges that describe where they're going.

Both views are on by default and can be toggled independently.

Available on every price chart: global and country-level, export, wholesale, and farm gate price types.

For the first time, Helios AI shows not just what climate models predict — but how accurate those predictions have been. Every climate chart now displays a Margin of Error, giving analysts a direct measure of forecast reliability before acting on the data.

Average absolute percentage point difference between Helios's 30-day WAPR forecasts and actual observed values — a clean ±% figure answering "how many percentage points off were we, on average?"

Updates weekly alongside regular pipeline runs. Available at global, country, and regional levels.

When Brazil soybeans show a WAPR of 20% with a Margin of Error of ±3%, you know Helios forecasts have historically landed within 3 percentage points of reality. No more guesswork about how much weight to give the forecast.

A new way to receive Helios intelligence without logging in. Weekly or monthly per-commodity summaries — delivered to your inbox — covering climate outlook, trade policy, geopolitical risk, yield and price forecasts, all in one place.

Format and cadence are still being refined through a live pilot. Open to both current customers and non-customers. If you're interested in receiving sample executive summaries, reach out to sales@helios.sc.

Last updated: May 2026 | Sabrina Lu, Head of Growth