54.4 GW
Vietnam peak demand · 2025 record
Asian Hackathon for Green Future 2026 · Top 30 of 439 · Renewable Energy & Low-Carbon Mobility
WE DON'T
PITCH BATTERIES.
WE GIVE THEM A BRAIN.
Internal war-room for Neural Grid Dispatch — the dispatch engine that makes Vietnam's battery fleet compliant, stable, and bankable. Everything we need to walk into VinUniversity and win.
Onsite finals beginJUL 2 · 2026 · HANOI
--
Days
--
Hrs
--
Min
--
Sec
10–16.3 GW
BESS needed by 2030 (Rev. PDP8)
<100 MW
BESS actually deployed today
0
Western dispatch rivals in Vietnam
01 / The Hackathon
The inaugural edition. No past winners to copy — we set the bar.
Run by Vingroup's "For Green Future" Foundation + VinUniversity + VinTechTalent, tied to the VinFuture Prize. 439 teams · 1,439 contestants · 22 countries. We're in the Top 30.
FORMAT
Train → Sprint → Cut
Online training (Jun 2–28) → onsite build sprint at VinUni → presentations → Top 9 advance to a final assessment round that picks the winner.
⚠ Build duration reported as both 24h and 36h — confirm.
PRIZES · $24,000 POOL
9 teams paid
1st $8,000
2nd $5,000
3rd ×2 $3,000
Consolation ×5 $1,000
+ mentorship & Vingroup-ecosystem network.
LOGISTICS
Travel + stay covered
All travel & accommodation for Top 30 fully covered for the Hanoi finals. Teams ≤4, current undergrad/master's students.
⚠ Visa support not stated — Pakistan team needs lead time. Ask early.
→ Critical Timeline
APR 6 – MAY 17, 2026 · done
Registration & preliminary submission
Idea proposal + intro video. Scored by the Technical Board.
EARLY JUNE 2026 · done
Top 30 announced — that's us ✓
Selected from 439 teams across 22 countries.
JUN 2 – JUN 28, 2026 · NOW
Online training phase — the prep runway
Intensive expert sessions + consultations on sustainability tech. Treat mentors as scouts: pressure-test our problem framing and demo plan here, before the clock starts. This is where we fix the proposal (see §05).
JUL 2 – JUL 5, 2026
Onsite hackathon + finals · VinUniversity, Hanoi
Timed build sprint → pitch to the panel → Top 9 → final assessment → winner. Working demo expected.
02 / The Panel
Five judges. Three lenses. Win all three.
A solution must be technically rigorous (AI/systems), commercially viable (entrepreneurship), and genuinely impactful (sustainability + corporate ESG). Tailor a line of the pitch to each chair.
Prof. Duong Nguyen Vu
CHAIR · VinUniversity Vice Provost · Chief Scientific Officer, AI Research Center · ex-NTU Singapore aerospace / air-traffic AI
How we win himLead with the AI: curtailment-probability forecasting, the cloud/edge split driven by physics, sub-100ms Rust control. He values rigorous, data-driven, systems-level thinking. Show the model, not just the story.
David Falcon
Director of ESG · VinFast (Vingroup's EV arm)
How we win himThis is our home judge — our track is his world. Frame NGD as enabling infrastructure for the green-mobility ecosystem (EVs need clean, stored, dispatchable power). Tie to net-zero-2050 and credible, deployable ESG impact — not vaporware.
Assoc. Prof. Carrie Ling
Director, Entrepreneurship Centre · Systems Integration & Design · HKUST
How we win herNail the business model: revenue stacking, <4.5yr payback, SaaS licensing path (Stage 3). Name the wedge clearly. She'll probe go-to-market and whether this is a company, not a project.
Dr. Paul Wang
Director, Technology Entrepreneurship Centre · Exec Director, Tech Entrepreneurship Academy · University of Hong Kong
How we win himSame as Ling — commercialization & scalability. Show the path from one pilot (Stage 1, historical data, no hardware) to pan-Asia SaaS. Be honest about staging and de-risking.
Assoc. Prof. Shauhrat S. Chopra
School of Energy & Environment · City University of Hong Kong · sustainability & life-cycle analysis
How we win himQuantify the environmental case conservatively: tonnes CO₂ avoided by displacing coal at the 6 PM peak, curtailment recovered, asset-life extension. He'll smell over-claimed impact — show methodology.
Organizing side — Toan Quach & Dr. Le Thai Ha
Marketing/CSR · "For Green Future" Foundation Managing Director
What they rewardStorytelling around impact. Quach is a Fulbright journalist; Le Thai Ha said winning ideas "arise from the insight, initiative, and determination of young people." A clear, human, well-told story travels — and gets amplified.
Panel sourced via official Vietnamese-language release — verify exact titles with organizers. No published numeric rubric exists; criteria are stated qualitatively as innovative · feasible · socially impactful · interdisciplinary.
03 / What Wins Here
VinFuture DNA: impact on millions, for the underserved, aligned to Vietnam.
The VinFuture Prize awards "science for humanity" — measurable impact at scale, Global-South equity, UN SDG alignment. 2023's Grand Prize went entirely to solar + battery tech. The judges think in that frame. So must our pitch.
PILLAR 01
Impact at scale
Lead with who benefits and how many, not the architecture. Quantify: GWh of curtailed solar recovered, tonnes CO₂ displaced, coal MW retired safely.
PILLAR 02
Feasible today
"Feasibility is non-negotiable." Our Stage 1 runs on historical grid data, zero hardware — provable in a weekend. Show the working demo path.
PILLAR 03
Vietnam-first, Asia-scale
Name net-zero-2050, JETP, PDP8 explicitly. Then show the SaaS replicates across ASEAN's emerging storage markets — mirroring Vingroup's own playbook.
The one-sentence pitch — hits every judge button
"Vietnam mandated 10–16.3 GW of battery storage by 2030 but has built under 100 MW — because batteries have no brain. Neural Grid Dispatch is that brain: it recovers curtailed solar, keeps the grid stable as coal retires, and makes storage bankable — turning Vietnam's net-zero pledge into deployable infrastructure, then scaling it across Asia."
science for humanity
net zero 2050
just transition
innovative · feasible · impactful
scalable across Asia
green generation
04 / Our Edge
We are positioned to win this track. Here's the unfair advantage.
LEAST CROWDED TRACK
~30% of teams
Renewable Energy & Low-Carbon Mobility is the least-subscribed of the three tracks (vs ~33% air quality, ~37% water/agriculture). Less noise, more signal — and it's exactly judge David Falcon's domain.
COMPETITIVE WHITE SPACE
No incumbent in Vietnam
Tesla Autobidder, Fluence, Wärtsilä GEMS — none operate in Vietnam. They're built for mature markets (NEM, CAISO, ERCOT). No identified local dispatch-software rival. We're building for the market Vietnam is creating right now.
PERFECT REGULATORY TIMING
Circular 62 — live Jan 26, 2026
Vietnam just passed its first standalone-BESS tariff (capacity + energy) — the first major ASEAN economy to do so. The revenue-stacking regime our product optimizes for didn't legally exist 6 months ago. We arrive exactly on time.
THE HARD PART IS THE MOAT
Temporal compliance ledger
The cryptographically-signed solar→battery→factory audit chain is the hardest, most defensible piece — and where our focus concentrates. Anyone can dispatch a battery; proving the green chain is the wedge.
05 / Fix Before We Pitch
Seven numbers in our proposal are wrong or risky. Judges include Vietnamese energy experts.
Our thesis is sound and current. These are precision fixes — but an expert catching one error costs us credibility on all of it. Correct them in the deck during the training phase.
1
Peak grid demand = 52.2 GW
→ Use ~54.4 GW (2025 record peak, +11% YoY).
52.2 is a 2023/early-2024 figure. Northern reserve margin is a critical 3–4% in dry season — a strong supporting point to add.
2
Coal = 58.7% of the daily mix
→ Reframe as a dry-season / Q1 2025 spike (~56.5%), not the annual average.
Full-year 2024 coal was ~50%. Stating 58.7% as the steady mix is the kind of thing Chopra or a Vietnamese expert will challenge. Say "up to ~56% in the dry season when hydro is low."
3
~300 MW of BESS deployed today
→ Under 100 MW is actually deployed. (300 MW was the OLD 2023 target.)
This one helps us — the real gap (<100 MW deployed vs 10–16.3 GW mandated) is even more dramatic. Largest real projects: EVN's 50 MW pilot, Marubeni–Vingroup 1.8 MW at VinPearl. Use "<100 MW."
4
Regulation mandates temporal matching / Green Labels / RECs
→ Frame as market & corporate-demand opportunity, not existing law.
Decree 57/2025 does NOT contain an explicit hourly time-matching mandate or a formal REC/Green-Label scheme. It's where the market is heading + what 24/7-CFE corporate buyers want. Present it as the demand we enable, not a legal requirement.
5
"Directive No. 10" demand response
→ Cite Directive 20/CT-TTg (target: cut peak load ≥1,500 MW via DSM/DR by 2025) + Electricity Law 61/2024.
"Directive No. 10" could not be verified as the DR instrument. Don't show an unverifiable number to expert judges.
6
Circular 62 governs frequency / ancillary payments
→ Circular 62/2025 = the standalone-BESS two-part tariff (capacity + energy). Frequency/ancillary pricing = Circular 11/2025 + Circular 21/2024 amendments.
Attribute each revenue stream to the right instrument. Getting the regulatory map exactly right is how we out-credential the room.
7
JETP = $15.5B flowing into the transition
→ $15.5B pledged but largely undisbursed; the US has withdrawn from the partnership.
Real as intent, unreliable as near-term capital. Don't lean on it as funding — our financial case stands on revenue stacking, not grants.
Sources: Vietnam Briefing (Revised PDP8 / Decree 57 / Circular 62), Energy-Storage.News (ASEAN BESS), Norton Rose Fulbright (Circular 62 / Decree 57), EVN/NSMO, GOV.UK (JETP). Full citations in research brief.
06 / The Pitch Playbook
A clean demo beats impressive tech that rambles. Every time.
Judges form an impression in 30 seconds and want the problem stated immediately. The demo is the largest single block — and the thing most likely to break. De-risk it like it will.
→ Pitch structure (~5 min · scale to slot)
0:00
Hook + Problem
One number. ~45s
0:45
Solution
What + who. ~30s
1:15
▶ DEMO
The centerpiece. ~120s
3:15
Market + Model
$ & scale. ~45s
4:00
Why us
Team. ~30s
4:30
Close + Ask
Direct. ~30s
MAKE THE INVISIBLE VISIBLE
NGD is middleware — so show data moving
- Build a thin dashboard. Curtailment forecast → pre-charge → discharge at 6 PM peak → signed compliance entry. Make judges watch the chain light up.
- Run on real historical grid data. Stage 1 is shadow-mode on historical data — that IS the demo. Replay a real curtailment day; show NGD capturing energy coal would've burned.
- Quantify the unseen. "This window: 4.2 MWh recovered, X tonnes CO₂ avoided, penalty avoided." Numbers turn infra into impact.
- Show the ledger entry. A real signed solar→battery→factory record on screen = the moat, made tangible.
DEMO DE-RISK CHECKLIST
Assume the wifi dies
- Pre-record the happy path. Clean screencast. If live fails: "let me show the recorded run" — never stall.
- Run everything locally / offline. Cache data, mock network calls, backend on localhost. Bring a hotspot.
- Pre-seed the data. Hit the impressive state instantly — no typing long inputs or cold model loads on stage.
- Bring the kit. USB-C→HDMI adapter, power strip, full battery + charger, backup laptop, notifications off, tabs pre-opened.
ROLES — ASSIGN BEFORE WE WALK ON
Presenter / domain — owns the problem story, model, and Q&A quarterbacking. Routes questions.
Builder(s) — drive the demo, own technical answers (Rust edge, forecasting, ledger).
Business lead — revenue stacking, payback, GTM, competitive answers.
The PM — guards scope and the clock; calls the "stop adding features" gate.
WHY STRONG TEAMS LOSE — DON'T DO THESE
Over-engineering / scope creep · no clearly defined problem · demo'ing slides instead of a working product · a broken demo with no fallback · not answering the rubric · unclear ask · one person hogging Q&A · dodging the competitor question (an instant-reject signal) · letting a well-meaning mentor steer a rebuild.
07 / Anticipated Judge Q&A
Rehearse the ten hardest questions. Out loud.
Q&A erases or cements the pitch. Never dodge competitors or risk. Answer the question actually asked. Pre-assign who speaks.
Q Who are your competitors, and why won't Tesla/Fluence just enter Vietnam? +
Name them directly — dodging this is an instant-reject signal. Tesla Autobidder, Fluence Mosaic, Wärtsilä GEMS are real and excellent — built for mature markets (Australia's NEM, CAISO, ERCOT) with established real-time price signals. Vietnam's market is being created right now (Circular 62, NSMO spun out 2024). Our wedge: purpose-built for Vietnam's two-part tariff, DPPA structures, and temporal-compliance needs that Western tools don't address. First-mover + regulatory fluency is the moat.
"We'd welcome them validating the market — by then we're the embedded compliance layer they'd integrate with."
Q Software with no hardware is useless — only <100 MW of BESS exists. What do you actually sell today? +
The Virtual Battery. Stage 1 needs zero physical BESS: we modulate non-critical industrial loads (HVAC, cooling) 15–20% via SCADA under Directive 20 demand-response to sustain temporal matching. The compliance + forecasting engine runs in shadow mode on historical data — provable now. Physical BESS dispatch is Stage 2. We sell analytics and compliance before a single battery is dispatched.
Q Is this a real company or a hackathon project? What's the business model? +
SaaS dispatch engine, B2B. Three stacked revenue streams compress BESS payback from 10+ years to under 4.5: (1) ancillary-services retainer (~15% capacity reserved for frequency control under Circular 11/2025), (2) penalty/REC-invalidation avoidance for factories (~20–30% of annual grid charges), (3) battery-life extension via managed depth-of-discharge (~7→12–15 yrs). We license to multinationals navigating DPPA compliance and to IPPs deploying under PDP8. Comparable models: Autobidder, Fluence IQ.
Q Coal is ~50% of the mix and provides grid inertia. How does software replace that? +
We don't replace the physics — we orchestrate the hardware that provides it. NGD manages grid-forming inverters on the BESS to emulate synchronous inertia and deliver primary frequency response. EVN itself cites lack of ancillary services to stabilise frequency/voltage as a key barrier to higher renewables. Until that gap is filled, the NSMO can't safely retire coal baseload — that's precisely the capability we provide.
Q What about three days of monsoon cloud cover? Batteries only hold 4–8 hours. +
Handled explicitly. NGD detects multi-day weather via satellite irradiance and triggers predictive hedging — tapping LNG balancing markets and hydro reserves to maintain continuity, without scope-creeping into carbon trading. We optimise within the grid's existing dispatchable reserves rather than pretending the battery alone covers a monsoon.
Q Cloud AI over 4G is too slow for grid frequency events. +
Correct — which is why we have a hard architectural split. Rust firmware on-premise at the edge handles all sub-100ms frequency physics with zero cloud dependency. The cloud handles 30-minute market/compliance decisions. The two communicate continuously but operate standalone during network loss. A grid event never waits for a round-trip.
Q Does Vietnamese law actually require this "temporal matching"? +
Be precise here (this is a credibility trap). No — it's not an explicit legal mandate yet. Decree 57/2025 enables DPPAs but doesn't mandate hourly matching or RECs. It's driven by corporate demand — multinationals with 24/7 carbon-free commitments need granular traceability to prove their Vietnam operations are genuinely green. We build the rails the market is moving toward, not a box-tick for a current rule.
Q Why blockchain? Sounds like buzzword bingo. +
Agree it shouldn't be the headline — it's an implementation detail, not the feature. The ledger (Hyperledger Fabric) is an append-only, tamper-evident store so a factory's solar→battery→consumption chain can't be retroactively edited when audited. If a simpler signed-log satisfies the auditor, we'd use that. The feature is verifiable green provenance; the tech is whatever proves it cheapest.
Q What's the environmental impact, quantified? +
Conservatively, per the curtailment data: Vietnam curtails 10–20% of solar at peak. Every MWh NGD shifts from noon-curtailment to the 6 PM coal-peak directly displaces coal generation. State it bottoms-up from one site, then scale — and flag the methodology rather than headline a giant number. Don't over-claim; Chopra will test it.
Q What's the ask? +
Frame it as catalytic: a pilot partnership with NSMO's test-bed and one industrial park to validate demand-response on live data — the one thing that unlocks Stage 2. For impact judges, the prize/network is what makes the pilot happen that otherwise wouldn't.
08 / Market Intel (verified)
The numbers, corrected and current. June 2026.
| Metric | Verified figure | Note for the deck |
|---|---|---|
| Peak demand (2025) | ~54.4 GW | Record; +11% YoY. North reserve margin 3–4% in dry season. |
| Total installed capacity | ~87.6 GW | Hit ~90 GW Nov 2025. |
| Renewables (solar+wind) | ~21.4 GW | ~26% of capacity; ~13–16% of generation. |
| Coal share | ~50% (2024) / ~56.5% (Q1'25) | Seasonal — frame the high number as dry-season. |
| BESS target by 2030 | 10,000–16,300 MW | Revised PDP8, Decision 768/QD-TTg (Apr 2025). ✓ accurate. |
| BESS deployed today | <100 MW | not 300 MW — that was the old target. |
| Solar curtailment | 10–20% at peak hours | Geographic mismatch: generation South, load North. |
| NSMO | Independent since Aug 2024 | Spun out of EVN; runs dispatch + wholesale market. |
| 500kV Circuit 3 | Live Aug 2024 | Doubled N–C transfer to 5,000 MW; eases not ends congestion. |
| Circular 62/2025 | Effective Jan 26, 2026 | First standalone-BESS two-part tariff in ASEAN. |
| Pre-reform payback | 10+ yrs (single stream) | Our wedge compresses it via stacking. |
RIVALS — NONE IN VIETNAM
Autobidder · Mosaic · GEMS
All built for NEM / CAISO / ERCOT. Vietnam is a genuine white space — state it as "no identified incumbent," not "none exists."
REGULATORY MAP
62 · 11 · 21 · 57
C62 = standalone BESS tariff. C11/2025 + C21/2024 = ancillary/frequency pricing. Decree 57/2025 = DPPA framework. Get each right.
NATIONAL FRAME
Net Zero 2050 · JETP · PDP8
PM Pham Minh Chinh's COP26 pledge. JETP $15.5B pledged (largely undisbursed, US withdrew). PDP8: coal → ~20% by 2030.
09 / Prep & Logistics
Energy management is a competitive edge. Show up ready.
BEFORE HANOI
During the training phase (now)
- Fix the 7 numbers in the deck (§05). Highest-leverage task this week.
- Lock the one-liner and the slide skeleton before any code.
- Build the demo data pipeline early — get a real historical curtailment day replayable.
- Use mentors as scouts — pressure-test problem framing + demo, but don't let them trigger a rebuild.
- Rehearse Q&A out loud (§07), assign who answers what.
PACK THE KIT
Onsite survival
- USB-C→HDMI adapter + power strip + 2 wall adapters + long charging cable.
- Personal hotspot + backup laptop — never trust venue wifi.
- Build-block schedule ~2hr blocks, breaks, a reset window every ~6hrs. Don't sprint a marathon.
- Fuel protein/nuts/water + hydration tabs — not sugar crashes.
- Travel docs passports, Vietnam e-visa lead time, meds, adapters.
10 / Ask The Organizers
Open questions — get answers before July.
These weren't published. Email the organizing committee (Toan Quach's team) during the training phase — knowing them is a quiet edge.
① Build duration
Is the onsite sprint 24h or 36h? Sources conflict — it changes our scope plan entirely.
② Pitch & demo format
Time limit per team? Slide count? Is a working prototype/demo mandatory or optional?
③ Scoring rubric
Is there a numeric rubric and weights? Only qualitative criteria are public.
④ Visa support
Do they issue invitation letters / assist visas for the Pakistan team? Biggest open logistics risk.
⑤ Post-win pathways
Any incubation, follow-on funding, or internships for winners beyond the cash?
⑥ Hour-by-hour agenda
Pitch order, arrival/departure, mentor access during the sprint.